Nectarios Ch. Benekos 1 , Rosy Nicolaidou 2 , Stathes Paganis 3 for the collaboration among:

1. l l l l Plans on H->4l (e,µ) analysis with CSC samples Nectarios Ch. Benekos1, Rosy Nicolaidou2, Stathes Paganis3 for the collaboration among: 1 Max-Planck-Institut für Physik, 2 Saclay muon group, 3 Sheffield University OUTLINE • Short reminder of what was done in the past • Full analyses on DC1 samples • Performance studies on Rome samples • Plans for present /future

2. Past Activities I What was done after the TDR (1999) using Data Challenge 1 samples • Full analysis chain with signal + background in all 3 channels • CBNT based analysis • TDR-like cuts • Main aim of these analyses was to do performance studies on lepton channels    ( e.g of talks given on the subject:Talk on Higgs Working Group25/5/2005, Talk on Physics Week 5/11/2004 ) ATL-COM-PHYS-2005-043 4µ 2e2µ • Resolution on mH : Worse by ~10% w.r.t TDR New geometry of µ-spectrometer (cracks) and more realistic simulation of ID (material + field)

3. +1% -1% eta eta Erec/Etrue eta eta 1TeV • Erec: new parameterization function introduced based on recent TB Analysis results (T.Carli et al.) • Extraction of new longitudinal weights • Become official after G4 migration • at “rome” samples individual weights for 3x5, 3x7, 5x5 clusters were calculated Past Activities I Linearity from 10GeV-1TeV

4. Past Activities I (Reconstruction efficiency in 4µ state) • At the TDR eff~84% for 4 ~16% inefficiency due to the signal acceptance • Now eff~65% for4 ~35% inefficiency out of which we know that ~16% signal acceptance ~ 9% due to new layout ~ 4% software problems ~ 6% (  6 ~ 1.5 % ) eff Muonboy for µ XKalman for ID 4 Single muons of Pt=20 GeV Combination ID-µ with STACO Order of magnitude of our incomprehension of the single-  efficiency Drop due to software problems Drop due to new layout ( crack)

5. Past Activities II – Rome samples What was done so far with the “rome” samples : • Migration of analyses in “eventView” framework and first tests • Stathes talk • event view wiki page • Performance studies with full pileup + cavern background (4µ channel) • Pile-up at low lumi 1033 • Cavern background with safety factor 1 • (sf011 x nominal ATLAS bkgr)

6. Past Activities II – Rome samples H (180 GeV) 4µ event view signal with full pileup + cavern background sample Combined ID-µ spectrometer efficiency using STACO package H(180)ZZ*4µsignal Efficiency for 4 µ ~81.3 % h H(180)ZZ*4µsignal +pileup +cavern bkgr Efficiency for 4 µ ~81% h

7. Plans towards CSC samples • Our Aim : • Work at Generator/filter level • Performance studies on CSC samples after each release • “where we stand” for a quick feedback • Full analysis on all 3 channels (4e, 2e2m,4m ) on a stable release (12.0.X ?) • 120 - 200 GeV Hmass region of interest • studies on miscalibrated and misalignement effects (detector “as-built”) • Use of the experience gained by CTB04 analyses (applied to e-id studies) • performance study of certain traditional e-id-cuts • studies on the systematic effects that each cut may involve • Method: Vary realistically the shapes of discriminant observables we are cutting on (e.g. shower width in the strips) and study the effect on the Higgs analysis • Comparison of these observables with the existing data to justify their use and performance. • identification of correlated e-id cuts so that we can remove those that are more sensitive to systematics than their correlated counterparts

8. Plans towards CSC samples (cont’d) • Generator level: • HiggsMultiLeptonFilter (in collaboration with D. Rebuzzi, S. Rosati, A. Nisati) • implementation of a Filter algorithm based on MultiLeptonFilter algorithm • cuts on invariant masses could be implemented to generate Zbb and tt backgrounds only in the signal phase space region • improve the rejection of background events that won’t be selected by the ATLAS Event Reconstruction and Analysis Package, while keeping acceptance 1 for the Higgs signal • performance studies as well as studies on possible biases introduced are still needed there • DC3 / CSC samples • in order to have a complete analysis with sufficient statistics (signal + bkgr) as it was done in DC1, the production of additional samples may be needed. • willing to participate, if needed (plans of the HiggsWG?) • The "complete analysis" is the key issue since a lot of things changed since DC1 • software chain • Detector geometry

9. Changes in the ATLAS Software since DC1 Major changes in ATLAS Software chain since DC1 in: • POOL/SEAL Event Persisency/Dictionary • Detector Geometry  GeoModel as the unique tool for all subdetectors • Geant4 simulation; Athena-based pile-up, digitization and event mixing • Full integration from RecoTaskForce designs/recommendation - new EDM • House-cleaning - robustness, performance, dependencies, etc. • Interactive as well as batch - Python job options files • GRID production - LCG2, NorduGrid, Grid3 • Testbeam-specific deliverables - DCS data, calibrations, real ByteStream, online monitoring, etc. • New Physics Analysis Tools • ….

10. ATLAS Geometry Layouts • Differences in DC3 and Rome with respect to DC2 Layout • Pixel: missing one layer • SCT: no difference • TRT: missing some end-cap wheels (C) • LAr: no difference • Tile: missing crack scintillators • Muon: initial layout Q instead of P • EEL and EES chambers removed • one layer of CSC removed • Switching between layouts is done by setting only one property in jobOpt.py file: • 'ATLAS-DC3-02' --> fixes to ATLAS-DC3-01 • 'ATLAS-DC3-01' --> 3 layers of pixel, no TRT-C wheels, • MuonLayout = R01.initial for the muon spectrometer which is essentially identical to Q02_initial_pro with the exception of additional, corrected descriptions of the passive material • 'Rome-Final' --> equivalent to the Rome2005 • 'Rome-Initial' --> equivalent to the Rome2005

11. EE chambers removed Increased radius of Barrel Half CSC NEW EC TOROID NEW FEET NEW BARREL TOROID Geometry layouts (Muon Spectrometer)

12. Detector Specifications for CSC Simulation • This includes accuracy of the geometry, plans for alignment studies and effects to include • in the digitization. Envelopes and Misalignments • In order to misalign the geometry in simulation it is necessary to leave enough room around • items that will be moved Summary of DD workshop on 15/11 and SPMB on 21/11

13. Analysis Tools • Event View Analysis framework • EventView allows us to run in parallel different combinations of the H->4l eventand easily study the performance of each of them • Example: The S/B for H->4l depends dramatically on the nature of the "electrons" and "muons" used. The best S/B may be coming from combination of egamma electrons and soft-electrons(muons). One may also want to keep particles which are outside the |eta|<2.5 ID region • Provides a clean/straightforward particle overlap removal • Provides a particle labeling scheme which allows the user to monitor all overlaps • Will provide: • generic fitting • (no more private Z->ee fitters), • multivariate and general statistical tools for advanced analysis • Allows us to create a miniAOD AthenaAware-ntuple with only the variables we are interested in including our own variables which are automatically written in the ntuple. • the ntuple allows back-navigation to the ESDs. • It can be read back in our H->4l analysis reducing tremendously the processing time • Repeat the old CBNT based analysis for cross-checks in the beginning at least 

14. Summary • Main aim is performance studies with the CSC samples • Lepton performance • Higgs performance and optimization • Usage of the new Physics Analysis Tools • EventView • Work onHiggs MultiLepton filter to optimize statistics of background samples

15. Geometry layouts (Inner Detector)